Solvent Chemistry in the Electronic Cigarette Reaction Vessel

Abstract

Knowledge of the mechanism of formation, levels and toxicological profiles of the chemical products in the aerosols (i.e., vapor plus particulate phases) of e-cigarettes is needed in order to better inform basic research as well as the general public, regulators, and industry. To date, studies of e-cigarette emissions have mainly focused on chromatographic techniques for quantifying and comparing the levels of selected e-cigarette aerosol components to those found in traditional cigarettes. E-cigarettes heat and aerosolize the solvents propylene glycol (PG) and glycerol (GLY), thereby affording unique product profiles as compared to traditional cigarettes. The chemical literature strongly suggests that there should be more compounds produced by PG and GLY than have been reported in e-cigarette aerosols to date. Herein we report an extensive investigation of the products derived from vaporizing PG and GLY under mild, single puff conditions. This has led to the discovery of several new compounds produced under vaping conditions. Prior reports on e-cigarette toxin production have emphasized temperature as the primary variable in solvent degradation. In the current study, the molecular pathways leading to enhanced PG/GLY reactivity are described, along with the most impactful chemical conditions promoting byproduct production.

Figure 1. Compounds identified herein by ¹H NMR in e-cigarette aerosols derived from a single puff from an electronic cigarette.

PG = propylene glycol; GLY = glycerol; 1a = propylene glycol hemiformal (major isomer); 1b = glycerol hemiformal (major isomer); 2 = glycidol; 3a = (Z)-prop-1-en-1-ol; 3b = (E)-prop-1-en-1-ol; 4 = dihydroxyacetone; 5 = acrolein; 6 = lactaldehyde; 7 = glycolaldehyde; 8 = glyceraldehyde; 9 = acetaldehyde; 10 = propanal; 11 = acetone; 12 = hydroxyacetone (acetol); 13 = acetic acid; 14 = formic acid; 15 = allyl alcohol.

Figure 2

(A) PG and GLY consumption as a function of device power (W) as well as puff duration (sec), determined by mass. (B) ¹H NMR spectra of unvaped control (bottom) PG/GLY solution along with aerosols derived from vaping PG/GLY at 4 W and 6 W showing product peaks increasing in number and intensity with increasing power.

Figure 3. ¹³C- and ²H-labeling study of model hemiformal production showing characteristic ¹H NMR properties consistent with the previously reported NMR spectra of 1a and 1b.

(A) CH₃OCH₂OH exhibits analogous peak positions as well as the characteristic O-H proton splitting pattern corresponding to those for 1a and 1b. (B) The ¹H-¹³C decoupled spectrum corresponds to that of the product formed by bubbling ¹²CH₂O in CH₃OH (spectrum C). (D) Upon addition of D₂O the hydroxyl resonance at 6.14 ppm diminished and the methylene protons at 4.53 ppm collapse to a singlet. (E) The full ¹H spectrum of CH₃OCH₂OH/CH₃OH in DMSO-d₆, for completeness.

Figure 4. Wattage-dependent product formation.

¹HNMR spectra were taken from a series of single puff samples of (PG/GL) collected in 1 W increments between 10W–15 W, showing the growing intensity of peaks associated with 2 (glycidol) and 3 (propanal enol isomers) as wattage is increased.

Figure 5. Expansion of the aldehyde region of a single puff-derived aerosol sample at 15 W.

The relatively large peaks corresponding to acrolein (5) are noteworthy. Compound 6 = lactaldehyde; 7 = glycolaldehyde; 8 = glyceraldehyde; 9 = acetaldehyde; 10 = propanal.

Figure 6. Aerobic Thermal Decomposition of Propylene Glycol.

Figure 7. Aerobic Thermal Decomposition of Glycerol.

Figure 8. Comparison of (PG/GLY) decomposition at power settings of 15 W using clearomizers of differing configurations from various sources.

An expanded region of the ¹H NMR spectra of aerosolized samples of PG/GLY in addition to an unaerosolized sample (bottom). The red spectrum (top) shows aerosol products generated via an inexpensive Sailebao^® CE4 cartomizer and shows the highest abundance and diversity of degradation products. Now considered an outdated design, the CE4 clearomizers were purchased as a component of a “Starter Kit” in 2014, and are still widely available. Samples collected from more sophisticated devices typically produce spectra containing fewer peaks of lower intensity, as demonstrated by the spectra from samples vaped using the KangerTech^® Protank-II and the Innokin^® iClear 16B, plotted in orange and green, respectively. Expensive, more current devices such as the Eleaf^® GS Air produce aerosolized samples showing relatively diminished product peak intensities as well as fewer degradation product peaks. While the extent of (PG/GLY) degradation varies between models, it is not unique to any one design.

Figure 9. Individual heating elements of the same design and manufacture, identical in appearance and packaged together as replacement units, can demonstrate wide variation in the abundance and profile of decomposition products observed in samples collected under controlled conditions.

Depicted above are results from aerosolized samples of PG/GLY collected under fixed conditions from the KangerTech^® Protank-II clearomizer, varying only the replaceable single-coil heating element. Three single-puff samples were collected at a modest power setting of 10 W from eight different replacement coils, three of which were labeled 2.2 Ω resistance by the manufacturer, while the other five were labeled 1.8 Ω resistance. (A) The intensity of NMR signals from several degradation products were compared by relative integration to the intensity of un-degraded PG/GLY peaks; the relative intensity of the thermal degradation products formic acid (Plot B, compound 14), hydroxyacetone (Plot C, compound 12), dihydroxyacetone (Plot D, compound 4), and glycidol (Plot E, compound 2) are plotted as percentages of the intensity of residual (PG/GLY). The average value among the eight replacement coils is plotted as a grey dashed line in plots B–E. All error bars denote a 90% confidence interval (n = 3). The asterisk (*) in plot C denotes p < 0.05 as determined by a two-variable, unpaired t-test.